Discrete Walrasian exchange process (Q1950353)

The paper under review shows that a discrete exchange Walrasian process attains a Pareto optimal allocation after a finite number of steps and the local equilibrium price supports the Pareto optimal allocation. The main contribution of the paper is to extend \textit{P. Champsaur} and \textit{B. Cornet}'s work [in: Economic decision-making: games, econometrics and optimisation, Contrib. in Honour of J. H. Drèze. Amsterdam: North Holland. 3--13 (1990; Zbl 0735.90014)] with each step endogenously determined by a local Walrasian equilibrium of an associated economy. The key feature of a local equilibrium is the uniqueness in terms of price and allocation. Section 2 presents a pure exchange economy with \(l\) commodities and \(m\) consumers. A \(\tau\)-local equilibrium of \(E(e) = (R_+^l, u_i, e_i)_{i=1}^m\) is a Walras equilibrium of the \(\tau\)-local economy \(E^{\tau}(e)\), where \(R_+^l\) stands for the consumption set, \(u_i\) is the utility function from \(R_+^l\) to \(R\) of the \(i\)-th consumer with initial endowment vector \(e\). A discrete Walrasian exchange process is a discrete dynamic system of \(\tau\)-local equilibrium \(W\). Proposition 2.2 shows that the utility level of each agent is nondecreasing, the allocation is individually rational and feasible, the exchanges are done according to the rate of exchange given by the price \(p(t)\) at each step, and the equilibrium point of the process is in the set of individual rational Pareto optimal allocations. The main result of the paper is Proposition 2.3 which shows that the discrete Walrasian exchange process attains an equilibrium point at a finite number of steps. The difference between this work and [loc. cit.] is that the discrete Walrasian process is a kind of discretization of the tangent economy process of Champsaur and Corneti. The advantage of this discretization comes from the uniqueness of local equilibria. The proof of this main result is occupied in Section 3. As the authors point out, they cannot expect to converge to a Walras equilibrium or to an element in the core since the consumers are myopic and the process does not keep track of the initial endowments. It remains to be shown that the discrete Walrasian process has an approximated neutrality property. The appendix contains proofs of two technical lemmas.